Statistics

In basketball (as well as all other team sports) statistics play an important role. They can show how good a player or a team is, give clues as to strengths and weaknesses of a team, and explain why a team won or lost a given game. However, they can often give misleading information if not looked at carefully.

3 rules of statistics

There are three basic rules of gathering and examining statistics: relative and absolute data, amount of data, and same unit calculations.

Relative and Absolute data

Although relative data can sometimes distort a picture, it can often clear it up. For example, take two imaginary players, Milton and Bradley. Milton has played in 20 games and scored 10 points in each game, for a total of 200 points. Bradley has played in 100 games and scored 2 points in each game, for a total of 200 points. Looking at their totals (200 points each), it seems that they are of about equal ability. However, it is easily seen that Milton is the far better player. That is why one should always use averages in similar calculations, so that it can be clearly seen who really is the better player.

Amount of data

You should always gather as much data as possible. The more data you have, the more accurate your results will be. For example, try clicking this button to flip the coin (if you can't do JavaScript, just flip a real coin and keep track of the results):
heads Heads: Tails: tails
There's no knowing what combination of heads and tails you got! The "expected" result is 5 heads and 5 tails, because there is equal probability of either happening. However, you may have gotten 6-4, 3-7, or any other combination. Remember: probability is chance, not laws. Now, try flipping the above coin 10 or 20 more times. Chances are that it's closer to the expected result. That is because the more data you have, the more accurate the result. In basketball, this rule also applies. For example, using the rating system described below, Michael Jordan (Chicago Bulls) has a rating of 33.4. Shaquille O'Neal (LA Lakers) has a rating of 42.2. Most people would agree that Shaquille O'Neal is not better than Michael Jordan, and even those who believe that he is better agree that he is not better by that much. A closer look at the statistics reminds us of several things. Below is a table of some important totals for these players during the 96-97 NBA regular season.
Name Games Points Rebounds Assists
Jordan, Michael 82 2431 482 352
O'Neal, Shaquille 51 1336 640 159

Obviously, since O'Neal played less games, his totals are less. However, if you find the averages for the players, a different picture forms. Below are the same stats, except that instead of totals, there are season averages in the points, rebounds, and assists column.
Name Games Points per game Rebounds per game Assists per game
Jordan, Michael 82 29.6 5.9 4.3
O'Neal Shaquille 51 26.1 12.5 3.1

These stats show that the two players are nearly equal in points and assists per game, but Jordan has a huge disadvantage in the rebounds column. Most basketball fans would notice that something is wrong. Shaquille O'Neal's scoring average (26.1) was very much affected by the several good games he had after his return. In this manner, his averages were affected more than they should have been. For example, in an NBA season our imaginary player Milton has a scoring average of 20.0 points per game (ppg) after playing in 81 of the 82 games of the regular season. That means that he scored a total of 1620 points this season. Our second imaginary player Bradley also has an average of 20.0 ppg, but he was injured half of the season and only played 40 games, which means that he has a total of 800 points. In the last (82nd) game of the season, the two players meet in a game and each one scores an incredible 60 points. After such a performance, Milton's average rises to 20.5 points. Bradley's average, however, rises to 21.0 points. This gives the impression that Bradley is much better than Milton, although they are almost exactly equal in ability. This shows that a good performance benefits a player who has played less games more than a player who has played more games, and very well illustrates the second rule of statistics: the more data you have, the more accurate the results. At first, it seems that this rule goes against rule #1 (take averages, not totals), but that is untrue. The best solution is to find the averages of Milton and Bradley over the past several years. That way, you will have an average representation, and you will receive the benefit of having more accurate data because it's "evened out" by the amount of data.

Same unit calculations

In areas such as basketball, one has many different statistics to deal with. Is the best player the one with the most points? The most rebounds? The most steals? In areas where there are many different types of data, it is important to convert all of the data to the same unit. Points, rebounds, assists, steals, turnovers, missed shots and fouls are some of the stats that each player has. For use on this site, we have developed a formula to give each player a rating, which may be used to compare him to other players and to find team ratings. This rating is a number indicating how many points is a player worth to his team. What follows is a step-by-step breakdown of each part of the formula.

Positives

Points-Since our unit of choice is points, we have to stick to it. Each point is worth one rating point.
Rebounds-An offensive rebound gives a good opportunity for scoring, and a defensive one takes away an opponent's chance of scoring. It's not worth 2 points because it's a chance that the opponent or you will score. It is not a 2 point basket that is being scored/taken away from opponent, it's a chance at one. Therefore, each rebound is worth 1.5 rating units.
Assists-An assist means that the player directly helped his team to score, so each assist is worth 2 rating points (a basket is usually worth 2 points).
Steals-A steal gives the team a chance to score, so it is worth 1.5 rating points.
Blocks-A block should be worth the same amount of points as a rebound. However, it should be remembered that unsuccessful blocks sometimes alter a shot a little, often enough for it to miss its mark. That is why blocks are worth 2 rating points.

Negatives

Turnovers-Just like a steal gains a chance at a shot, a turnover gives it up. Therefore, turnovers are worth -1.5 rating points.
Personal Fouls-A foul not only gives the other team a possible chance to shoot foul shots, 6 fouls in one game mean that the player is disqualified for the rest of the game. Also, most coaches will not keep a player with 4 or 5 fouls in the game until the last minute, for fear that the player may get another foul and be disqualified. That is why a foul is worth -2 rating points.
Shots Missed-A missed shot gives the other team a chance for a rebound and score, as well as not scoring. However, offensive rebounds make up for that miss. Each missed shot that is not rebounded by the same player has a 50-50 chance of being rebounded by either team. So then the difference [Missed Shots-Offensive Rebounds] is worth -1 ratings point.

Final Rating Formula

So, after all the factors are added/subtracted, the final formula is:

{ [(points)+(rebounds*1.5)+(Assists*2)+(Steals*1.5)+(Blocks*2)]-[(Turnovers*1.5)+(Fouls*2)+(Shots Missed-Offensive Rebounds)] }/Games Played

Now that you know all the principles of statistics, why don't you pick two (W)NBA teams and have some fun seeing who wins, in our Game Predicter game. But remember, the results given are far from the truth, because of the fact that team ratings are simply the average of the players on the team, and don't in any way reflect reality, because of injuries and the above-described factors.

[an error occurred while processing this directive]